Skip to content

Ask your warehouse. Get an answer you can check.

Connect read-only views, ask in plain language and choose the shape of the answer. exelens loads only what the question needs, checks every figure against the rows it loaded, and lets you pin the result to a board that keeps itself current.

Start free trialSign in

14 days free with no card. Bring your own AI key.

Question:

Bar chart

West led Q3 net revenue at $1.42M, 31% of the total.

4 of 4 figures match the loaded rows

SELECT "region", "net_total", "ordered_at" FROM "analytics"."v_orders_daily" WHERE "ordered_at" >= :period_start AND "ordered_at" < :period_end

Try another

Example answers on sample data.

Your warehouse in. Your AI on top. exelens out.

exelens sits between the data you already model and the AI you already pay for. Read-only views go in, one checked visual comes out, and the session disappears.

Your data

Read-only views, as you run them today

  • Snowflake
  • BigQuery
  • Databricks
  • Amazon Redshift
  • Azure Synapse
  • Microsoft Fabric
  • PostgreSQL
exelens

Your AI

On your own key, no markup on tokens

  • Anthropic
  • OpenAI
  • Google Gemini
  • xAI
  • Amazon Bedrock

Business questions change every day. Boards should too.

From a view to a board in three steps

  1. 1

    Connect read-only views

    Give exelens a read-only identity and import only the views you want answered from. It warns you if the identity can write, and checks every view before anyone asks.

  2. 2

    Ask, and choose the output

    Every question comes with its shape: a KPI, a bar or line chart, a table or a written summary. The answer arrives in that shape with its assumptions and the exact statements it ran.

  3. 3

    Pin it to a board

    A pinned answer becomes a graph that refreshes on a schedule. Data-only refresh reruns the same statements with no AI cost, and alarms tell you when a number crosses a line.

exelens is only as sharp as the views behind it

It answers from views your team already curated, so every figure traces back to a statement a person wrote. The better the views, the better the answers.

What “ready” looks like

  • Curated views

    Marts that already join and clean. Not raw tables.

  • One grain, one date

    Each view says what a row is, and which column is time.

  • Described views

    Names, time column and meaning sit on the view, not in every question.

  • Safe to expose

    Read-only identity. Only the views you want answered from.

A model cannot guess which orders count as revenue. Curated views are the work that comes first.

What exelens is for

Seven things the product actually does. Not a feature list for a brochure: the reasons a team would switch.

Security and privacy

Your data never leaves your warehouse. The model never writes SQL.

exelens asks for a read-only identity, imports only views, and encrypts every credential with a per-workspace key. Loaded rows expire with the session. PII can be masked or hidden. Business adds SSO, SCIM and an IP allowlist.

  • No AI-written SQL

    The model never authors a statement. We run one fixed, verified read of a view you already own.

  • Nothing sensitive at rest

    Raw view rows live in a short session and are deleted. Pins keep the chart, not the warehouse.

  • Your warehouse, your key

    Compute stays where the data is. AI is billed by your provider, with no markup on tokens.

  • Encryption, SSO, PII

    Per-workspace keys, SSO and SCIM on Business, column masking. A read-only identity, every time.

Your views

The logic already lives in the warehouse.

Import only the views you want answered from. Descriptions, time columns and column visibility sit on those views. The model reads your views as you wrote them; it does not invent SQL or import tables.

  • Views only, never tables
  • Time column and descriptions
  • Hidden and masked columns
  • Checked on import

Ephemeral BI

Traditional reports go stale. This board is for the question you have today.

The usual stack

An analyst builds a dashboard for last quarter’s question. The business moves. The tile is still there in six months, unopened, quietly wrong.

Here

You ask, you pin, you refresh. Tomorrow’s question replaces today’s visual. Idle graphs get flagged. Nothing is meant to live forever.

Multi-platform

Not locked to Cortex. Not locked to Genie.

Warehouse-native chat answers only inside that vendor. Most companies run more than one system. A workspace in exelens can mix views from Snowflake, BigQuery, Databricks, Redshift, Fabric and Postgres, and one chat can switch Anthropic, OpenAI, Gemini, xAI or Bedrock on the keys you already pay for.

A fraction of the cost

Workspace price, not a seat tax on everyone who looks at a number.

Illustrative list prices, USD, 2026. Competitors change packs often; this is the shape of the bill, not a quote.

ToolTypical listWhat you actually buy
Power BI$10–24 / userPlus Fabric capacity as you grow. Viewers are paid seats.
Tableau$15–75 / userCreator, Explorer, Viewer. A five-analyst team is already hundreds a month.
Cortex / GenieInside the platform billNo extra SaaS line. Locked to one warehouse and that vendor’s models.
exelens$59–599 / workspaceViewers free on Team and above. Your AI key, no token markup.

Five builders and a dozen executives: Tableau or Power BI charge the viewers. Cortex and Genie hide the cost in a warehouse contract you cannot take to another vendor. Team is $249 a month, viewers free, AI on your own key.

For everyone

Analysts, managers, the C-suite. Connect the warehouse, paste the key, ask.

No desktop install, no six-month BI program. A stranger can sign up, attach a read-only view and an AI key, and have a checked visual in minutes. The subscription is the workspace. People who only read boards do not need a paid seat on Team and above.

Not a generic LLM

One question, one visual. Nothing else.

ChatGPT will write a poem about your revenue. exelens will not. Every message names an output type first. Off-topic, injection and “just write SQL” are refused. The model exists to produce a pinnable graph the business can use today, then let it go.

Pin the answer. It keeps itself current.

Scheduled refresh
Hourly to monthly, in your workspace's time zone and fiscal calendar.
Data-only or AI refresh
Rerun the saved statements for free, or let the model re-read new data and rewrite the summary.
Alarms
Set a threshold or a change on any graph and get an email, Slack or Teams message.
Share without seats
Free viewers, guest access by email and password-protected snapshot links that expire.

Plans that start small

Every plan includes the whole product. Limits grow with your team; your AI spend stays on your own key.

Trial

14 days, no card

$0for 14 days

  • Every feature to try it on your own data
Questions a month
100
Graphs
10
Boards
1
Exports
csv
Start free trial

Solo

For one analyst or founder

$49per month, billed yearly

  • 1 seat, 2 connections
  • 1,000 questions a month
  • Scheduled refresh
Questions a month
1,000
Graphs
30
Boards
3
Exports
csv, xlsx
Choose Solo

Team

Most teams start here

For a data team and its executives

$199per month, billed yearly

  • 5 seats, unlimited free viewers
  • Guests, snapshot links and digests
  • PDF and PowerPoint export
Questions a month
5,000
Graphs
150
Boards
20
Exports
csv, xlsx, pdf, pptx, png, svg
Choose Team

Business

For companies running on boards

$499per month, billed yearly

  • 15 seats, SSO
  • Graph alarms, Slack and Teams
  • Full REST API and MCP endpoint
Questions a month
25,000
Graphs
600
Boards
100
Exports
csv, xlsx, pdf, pptx, png, svg, parquet
Choose Business

Enterprise

Custom limits. Private connectivity. Security review and SLA.

Talk to us

Prices in USD, before tax. AI usage is billed by your own provider on your own key, so there is no markup on tokens. Viewers who only read boards are free on Team and above.

Why exelens exists

Dashboards answer yesterday’s questions. Chat tools answer today’s, but hand a model your database. We wanted both: the speed of asking, and the evidence of a report.

The name

exelens

Short for executive lens

The product is a lens for the people who have to stand behind the number: one question, one visual they can check, then pin or let go. “Executive lens” was accurate and too long. We kept the two halves that matter, exe and lens, so the mark itself is the idea. Not a chatbot. Not another dashboard suite. A way to look at the business today.

Who it is for

  • Data teams

    You own the views. The rest of the company can use them without opening a ticket.

  • Executives

    A number before the meeting, with enough evidence to stand behind it.

  • Growing companies

    Past spreadsheets, not ready for a six-month BI program. A few views and your own key are enough.

The common thread is a warehouse with views already modeled. Not there yet? See what ready looks like.

Why you should care

  • Minutes, not a queue

    The question that waited for an analyst gets a visual while the meeting is still going.

  • Numbers you can defend

    Every figure is matched to the rows it came from, or withheld.

  • Lower risk

    The model never writes SQL. Loaded rows disappear when the session ends.

  • $

    Honest costs

    AI billed by your provider. People who only read boards are free on Team and above.

Why exelens and not something else

Classic BI

Speed
Someone builds a dashboard first.
SQL
An analyst, once per report.
Trust
Hope the author got it right.
Your data
Often copied into the tool.

Chat with your data

Speed
Fast, in free-form chat.
SQL
The model writes it.
Trust
Read the SQL yourself.
Your data
Often copied. AI billed in.

exelens

Speed
Fast. One visual, the type you chose.
SQL
A fixed, verified read of your view.
Trust
Matched to the rows, or withheld.
Your data
Stays in your warehouse.

Built by Zenot: data platforms and production AI, every day. That is why the product is strict about what it will answer, and why we can help with the work that comes before it.

Questions people ask first

Do you store my warehouse data?

Loaded rows live in a short-lived working session and are deleted when it ends. When you pin an answer, exelens keeps the result of that answer (not the source rows) so the board can show it.

Can the AI change or delete anything?

No. The model chooses from a fixed set of read-only tools. exelens builds one kind of statement, a SELECT of listed columns from one imported view for one period, verifies it, and runs it with a read-only identity.

Which AI models can I use?

Models from Anthropic, OpenAI, Google, xAI and Amazon Bedrock, on your own key. You choose a default model per workspace and can switch per question.

What happens when a question can't be answered?

exelens tells you before loading data: which view or column is missing, or what to ask instead. If figures in an answer can't be verified, it withholds them rather than guessing.

Who can see a board?

Boards are private until you share them. Share with your workspace, invite guests by email, or create a password-protected snapshot link with an expiry date.

What if our views are not curated yet?

exelens answers from views your team has already modeled. If you are not there yet, Zenot, the team that built it, designs that foundation. Talk to Zenot.

What does the trial include?

Every feature for 14 days on one connection, with no card. Your data and boards carry over when you choose a plan.

Bring one view and one question.

Start free trialSign in